Abstract
It is increasingly recognized that measures of quality of life (QL) need to be integrated in assessments of outcome following treatment for childhood cancer. Requirements for such scales are discussed with respect to four issues: the need to be (i) brief and comprehensive, (ii) reliable and valid, (iii) able to account for developmental changes in meaning of QL and (iv) amenable to proxy ratings. Currently available instruments are reviewed and their advantages and disadvantages considered. In developing new instruments, attention needs to be paid to differences in definitions of QL made by clinicians and families. New instruments must take more account of patient and family perspectives. More multicentre work is called for in order to recruit samples of sufficient size that acceptable levels of reliability and validity can be established.